Upgrading SMB packages

Use these steps to upgrade IBM Spectrum Scale SMB packages.

Attention: All protocols must also be upgraded to the newer code level along with base GPFS.

SMB package update in an IBM Spectrum Scale cluster is done in two phases to minimize the downtime of the SMB service. The protocol nodes in the cluster are divided in two halves and SMB is updated on each half one by one. In the following steps, NodeList1 is the comma-separated list of nodes in the first half of the cluster and NodeList2 is the comma-separated list of nodes in the second half of the cluster.

Note: All protocol nodes that are running the SMB service must have the same version of gpfs.smb installed at any time. A brief outage of the SMB service is required to upgrade gpfs.smb to the newer version across all protocol nodes. The procedure that is outlined here is intended to reduce the outage to a minimum.
  1. On the first half of the protocol nodes, do the following steps.
    1. Issue the following command to suspend the nodes and stop all the protocol services that are running on these nodes.
      mmces node suspend -N NodeList1 --stop
      Note: Suspending nodes triggers IP address reassignment and client failover. For more information, see SMB fail-over scenarios and upgrade.
    2. On each node in the first half of protocol nodes, issue one of the following commands, depending on the operating system, to upgrade the SMB package.
      • Red Hat Enterprise Linux or SLES 12
        rpm -Uvh gpfs.smb*.rpm
      • Ubuntu 16.04 LTS or 18.04 LTS
        apt install ./gpfs.smb_*.deb

        The name of the IBM Spectrum Scale SMB package for Ubuntu 16.04 LTS is in this format: gpfs.smb_version_string_architecture.deb.

        The name of the IBM Spectrum Scale SMB package for Ubuntu 18.04 LTS is in this format: gpfs.smb_version_string~bionic_architecture.deb.
        Note: If you upgraded the operating system on the protocol nodes from Ubuntu 16.04 to Ubuntu 18.04, uninstall any obsolete packages by issuing the following command on each of these nodes.
        apt-get autoremove
  2. On the second half of the protocol nodes, do the following steps.
    1. Issue the following command to suspend the nodes and stop all the protocol services that are running on these nodes.
      mmces node suspend -N NodeList2 --stop
      Note: Suspending nodes triggers IP address reassignment and client failover. For more information, see SMB fail-over scenarios and upgrade.
  3. On the first half of the protocol nodes, issue the following command to resume the nodes and start all the protocol services that are enabled on these nodes.
    mmces node resume -N NodeList1 --start

    For information on viewing the health status of a node, and detailed information about the health status and potential corrective actions, see System health monitoring use cases and Events.

  4. On the second half of the protocol nodes, do the following steps.
    1. On each node in the second half of protocol nodes, issue one of the following commands, depending on the operating system, to upgrade the SMB package.
      • Red Hat Enterprise Linux or SLES 12
        rpm -Uvh gpfs.smb*.rpm
      • Ubuntu 16.04 LTS or 18.04 LTS
        apt install ./gpfs.smb_*.deb

        The name of the IBM Spectrum Scale SMB package for Ubuntu 16.04 LTS is in this format: gpfs.smb_version_string_architecture.deb.

        The name of the IBM Spectrum Scale SMB package for Ubuntu 18.04 LTS is in this format: gpfs.smb_version_string~bionic_architecture.deb.
        Note: If you upgraded the operating system on the protocol nodes from Ubuntu 16.04 to Ubuntu 18.04, uninstall any obsolete packages by issuing the following command on each of these nodes.
        apt-get autoremove
    2. Issue the following command to resume the nodes and start all the protocol services that are enabled on these nodes.
      mmces node resume -N NodeList2 --start

      For information on viewing the health status of a node, and detailed information about the health status and potential corrective actions, see System health monitoring use cases and Events.